improve container interconnectability

[the-leonat]

At the moment one container can connect to the other (if on the same network) by assigning a static ip when attaching the container to the docker network with the --ip custom docker option.
Rolling deployments are not longer possible and thus if the deployment failes the service is not available, which is not good at all.
But it makes sense as containers during deployment need to coexist (the old running one, the new one to deploy) in order for rolling deployments to work.
I suggest to implement the docker run parameter --network-alias which would resolve this issue, as multiple containers can have the same alias on the same network. We could then reference containers by their network alias instead of their internal ip respectively.

[travistylervii]

Yes! this is needed.

[matthieu-beteille]

+1

[the-leonat]

just wanted to check in, if this will be implemented anytime soon. @andrasbacsai

[stevenirby]

Maybe an environmental variable could be used for this? I would take that over how it works now.